Abstract
The Arctic Ocean is particularly vulnerable to ocean acidification, a process that is mainly driven by the uptake of anthropogenic carbon (C(ant)) from the atmosphere. Although C(ant) concentrations cannot be measured directly in the ocean, they have been estimated using data-based methods such as the transient time distribution (TTD) approach, which characterizes the ventilation of water masses with inert transient tracers, such as CFC-12. Here, we evaluate the TTD approach in the Arctic Ocean using an eddying ocean model as a test bed. When the TTD approach is applied to simulated CFC-12 in that model, it underestimates the same model's directly simulated C(ant) concentrations by up to 12%, a bias that stems from its idealized assumption of gas equilibrium between atmosphere and surface water, both for CFC-12 and anthropogenic CO(2). Unlike the idealized assumption, the simulated partial pressure of CFC-12 (pCFC-12) in Arctic surface waters is undersaturated relative to that in the atmosphere in regions and times of deep-water formation, while the simulated equivalent for C(ant) is supersaturated. After accounting for the TTD approach's negative bias, the total amount of C(ant) in the Arctic Ocean in 2005 increases by 8% to 3.3 ± 0.3 Pg C. By combining the adjusted TTD approach with scenarios of future atmospheric CO(2), it is estimated that all Arctic waters, from surface to depth, would become corrosive to aragonite by the middle of the next century even if atmospheric CO(2) could be stabilized at 540 ppm.